grill-me-M
Skill1 starsChanged 4 months agoMITView on GitHub
Interview the user relentlessly about a plan, design, or topic until reaching shared understanding. Walks each branch of the decision tree. Decisions are captured in a distilled brainstorm file (delegated to notes-M). Use when user wants to stress-test a plan, get grilled, run a discovery session, or says "grill me".
In this repository6 other files it gives its agents.
- handoff-MSkill
- html-report-MSkill
- notes-MSkill
- pre-compactSkill
- skill-MSkill
- sql-style-MSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 4 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 13
All file operations (create / reopen / structure / checkpoint / curation / status / reconcile) live in the notes-M skill.
- Distinctive: line 29
Ask one question at a time.
- Distinctive: line 31
If a question can be answered by reading a file/doc or exploring the codebase, do that instead of asking.
- Distinctive: line 32
When I can't answer something, capture it as a flag (item → owner) in the notes file and move on.
- Distinctive: line 33
Keep going until I say we're done or you've covered every branch.
- Very specific: line 41
When the interview reaches a conclusion (or is being parked), trigger notes-M's end-of-topic reconcile: read the file for contradictions, …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Grill Me
Relentlessly interview me about every aspect of the topic until we reach shared understanding. Walk down each branch of the decision tree, resolving dependencies one by one. Goal: extract what's in my head into a durable, organized file.
File mechanics — delegated to notes-M
All file operations (create / reopen / structure / checkpoint / curation / status / reconcile) live in the notes-M skill. Use it in interview mode.
notes-M will auto-invoke when it detects a multi-step interview. You don't need to call it explicitly — just follow the interview discipline below, and write each captured decision to the file notes-M creates.
Body section in the file = ## Decisions log (see notes-M for the consolidated-entry format: Decision · Why · Dead-ends · Flags).
Question style
For each question:
- Present numbered or lettered options whenever possible so I can reply with just a number, letter, or one word.
- For open-ended questions that can't be simplified, mark them clearly as needing more detail.
- Provide your recommended answer (your best inference from context) so I can just confirm, correct, or redirect.
- Minimize my typing — aim for replies of one word, one number, or 1–2 sentences max.
Question flow
- Ask one question at a time. Wait for my answer before moving on.
- Resolve dependencies in order: settle the upstream decision before ones that depend on it.
- If a question can be answered by reading a file/doc or exploring the codebase, do that instead of asking. If I hand you a doc (Google Doc, etc.), read it and surface only what's net-new.
- When I can't answer something, capture it as a flag (item → owner) in the notes file and move on. Don't stall.
- Keep going until I say we're done or you've covered every branch. Offer a completeness backstop near the end ("anything we haven't touched?").
Checkpoint after every answer
Non-negotiable: after EVERY answer, BEFORE the next question, update the notes-M file with the new decision/flag. Apply notes-M's curation rules (collapse, supersession, pivot). Never batch.
At the end
When the interview reaches a conclusion (or is being parked), trigger notes-M's end-of-topic reconcile: read the file for contradictions, fix them, set status to concluded or parked, give me a short recap.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

